Rifapentine is a promising agent for the treatment of tuberculosis, combining potency and prolonged serum concentrations above the minimal inhibitory concentration (MIC) of Mycobacterium tuberclosis after a single dose. These factors led to the evaluation of rifapentine in once-weekly regiments for tuberculosis. However, in an ongoing CDC sponsored trial the once-weekly regimen of rifapentine and isoniazid had a higher relapse rate than the control regimen of twice-weekly rifampin and isonizid. The higher relapse rate may be due to an unherent lack of efficacy of weekly therapy, but may also be due to poor bioavailability of the riapentine formulation used in the trial. It has long been known that differences in drug formulation can dramatically affect the absorption of rifampin, and the same may be true of rifapentine. The rifapentine formulation used in the CDC-sponsored rifapentine trial had adequate bioavailability in studies using normal volunteers, but this has not been documented in HIV-infected patients or among patients being treated in the anti-tuberculosis trial. There is growing evidence that HIV infection is associated with the occurence of rifamycin-resistant tuberculosis (TB). Rifampin mono-resistant tuberculosis among HIV-infected persons has been documented in several areas of the United States and does not appear to represent the widespread dissemination of a small number of drug-resistant strains. Emergence of rifampin resistance during directly observed therapy was recently documented in several patients with AIDS. A report from San Francisco documented an association between HIV infection and acquisition of drug resistance. Perhaps the most convincing evidence of the association between HIV infection and rifamycin resistance was the occurrence of rifampin-resistant relapse among 4 of the 36 HIV-infected patients randomized to receive once-weekly rifapentine plus isoniazed. whereas this has not occurred among HIV-infected patients randomized to twice-weekly rifampin plus isoniazid or HIV-negative patients randomized to either treatment regimen. These studies suggest that there may be a marked propensity for abnormal pharmacokinetics of TB drugs in patients with HIV infection. This study will investigate the possibility.